When moving house, we have to sort our belongings into boxes and transport these boxes to our new abode — a function that the GGA (Golgi-localized, γ-ear-containing, ADP-ribosylation factor (Arf)-binding) proteins seem to have in the cell. GGAs associate with the cytosolic face of the trans-Golgi network (TGN) and sort mannose 6-phosphate receptors (MPRs) into clathrin- and GGA-coated carriers, which are then transported to endosomes. The interactions that deliver these carriers to endosomes have remained unclear, but now, reporting in The EMBO Journal, Bonifacino and colleagues provide new insights.

The monomeric GGA proteins contain four domains: a VHS domain (which binds MPRs); a GAT domain (which binds activated Arf-family GTP-binding proteins); a hinge region (which binds clathrin); and a GAE domain, which is thought to bind accessory factors that can regulate the function of GGA-containing coats or GGA-coated carriers.

Bonifacino and co-workers began by looking at the interaction of the human GGAs (GGA1, GGA2 and GGA3) with Rabaptin-5, a protein that is important in endosome fusion. Although previous studies had shown that GGA-GAE–Rabaptin-5 interactions were weak, and therefore probably not physiologically significant, the authors decided to investigate these interactions further because of their possible implications.

By carrying out pull-down experiments using various glutathione-S-transferase (GST)–GGA domains and Rabaptin-5 from bovine brain cytosol, the authors found that GGA-GAE domains interact with Rabaptin-5 in the Rabaptin-5–Rabex-5 complex (a complex that regulates endosome fusion), and that these interactions are stronger than was previously thought. They confirmed that the GGA–Rabaptin-5 interaction occurs in vivo by immunoprecipitating endogenous proteins.

The authors then used the yeast two-hybrid system to analyse the structural determinants of the GGA–Rabaptin-5 interaction. They showed that residues 428–455 of Rabaptin-5, which are in a predicted random coil, contain the minimal sequence needed for interactions with the GAE domains of GGA1–3. In addition, they showed that the GGA–Rabaptin-5 interaction is bipartite for GGA1 and GGA2, as the carboxy-terminal coiled-coils of Rabaptin-5 interact with the GAT domains of these GGAs.

Using alanine-scanning mutagenesis, Bonifacino and colleagues were able to further define the putative GGA-GAE-binding motif in Rabaptin-5. This motif is FGXLV from residues 439–443, where X is any amino acid.

So, what happens when Rabaptin-5 binds GGAs? The authors studied the effect of His6-tagged Rabaptin-5 fragments on the binding of clathrin to GST–GGA-hinge+GAE constructs in vitro, and found that Rabaptin-5 interferes with clathrin–GGA interactions.

Finally, using immunofluorescence microscopy, the authors studied the change in the localization of GGA1 and its associated MPR cargo in HeLa cells after these cells had been transfected with green fluorescent protein–Rabaptin-5. Before transfection, they found that GGA1 and its cargo were localized to the TGN, whereas after transfection, endogenous GGA1 and MPR were localized to Rabaptin-5-stabilized large endosomes.

These data have revealed “...a functional link between proteins regulating TGN cargo export and endosomal tethering/fusion events”. They have also allowed Bonifacino and co-workers to suggest that GGA–Rabaptin-5–Rabex-5 interactions cause clathrin to be released from GGA-coated intermediates or prevent clathrin re-binding, which might allow TGN-derived carriers to fuse with endosomes and deliver their MPR cargo.